Helicobacter hepaticus infection in mice: Models for understanding lower bowel inflammation and cancer

Abstract

Pioneering work in the 1990’s first linked a novel microaerobic bacterium, Helicobacter hepaticus, with active chronic hepatitis and IBD in several murine models. Targeted H. hepaticus infection experiments subsequently demonstrated its ability to induce colitis, colorectal cancer, and extra-intestinal diseases in a number of mouse strains with defects in immune function and/or regulation. Helicobacter hepaticus is now widely utilized as a model system to dissect how intestinal microbiota interact with the host to produce both inflammatory and tolerogenic responses. This model has been used to make important advances in understanding factors that regulate both acquired and innate immune response within the intestine. Further, it has been an effective tool to help define the function of regulatory T cells, including their ability to directly inhibit the innate inflammatory response to gut microbiota. The complete genomic sequence of H. hepaticus has advanced the identification of several virulence factors and aided in the elucidation of H. hepaticus pathogenesis. Delineating targets of H. hepaticus virulence factors could facilitate novel approaches to treating microbially induced lower bowel inflammatory diseases.

Initial isolation and characterization of H. hepaticus and H. bilis

Since the isolation and characterization of Helicobacter pylori, the etiologic agent of peptic ulcer disease in 1983, a number of other Helicobacter species have been isolated, many of which can produce serious non-gastric disease in their respective hosts. Helicobacter species that colonize the lower bowel and biliary tract of mice have been associated with the development of colitis resembling human inflammatory bowel disease (IBD) in susceptible hosts. Further, H. hepaticus infection has been demonstrated to exacerbate the development of cancer at both intestinal and extraintestinal sites. The most well studied member of this group of enterohepatic Helicobacter species (EHS) is Helicobacter hepaticus. In this review we will focus on the initial identification of H. hepaticus, the basis for H. hepaticus-induced lower bowel inflammation, and the characterization of bacterial virulence factors.

In the early 1990’s, an active chronic hepatitis was detected in several inbred strains of mice including A/JCr, originating from a barrier-maintained facility. A distinctive morphological pattern of liver damage was noted which had not been previously identified in mice. Using a special silver stain, a spiral-like organism was observed in the liver. Infected A/JCr mice were then shipped to our laboratory for characterization and diagnostic evaluation. A microaerobic bacterium was consistently isolated from the liver and colon of these mice between 3 and 18 months of age. By biochemical and molecular methodologies these bacteria were confirmed to represent a new Helicobacter sp. which we named H. hepaticus ^1-3. Shortly after the isolation of H. hepaticus, we isolated a fusiform bacterium from the liver, bile, and lower intestine of aged, inbred mice. On the basis of 16S rRNA gene sequence analysis, the organism was classified as a novel Helicobacter sp., H. bilis.

Currently there are 9 formally named EHS isolated from mice as well as several other novel Helicobacter spp. awaiting formal naming. Most recently, H. pullorum, a human pathogen, has been isolated from commercial, barrier maintained mice ⁴. Surveys conducted in 1995 and 2007 indicated that H. hepaticus as well as other EHS have a worldwide distribution in academic and commercial mice ^{5, 6}. Contrary to the earlier survey, no Helicobacter spp., however, were found in mice in 2007 from advertised Helicobacter-free production areas from two U.S. vendors. These results demonstrated that infection with EHS is wide spread in most if not all mouse colonies used for experimental investigations.

Spontaneous IBD in immunodeficient and genetically engineered mice: Association with Helicobacter spp

Prior to the isolation of EHS, it had been speculated that microflora might play an important role in the etiology of spontaneous colitis that developed in a number of immunodeficient laboratory mouse strains including IL-10-deficient, scid, and RAG. Berg et al. also demonstrated that inbred mice on different strain backgrounds with undefined intestinal flora influenced the degree of colitis and lower bowel adenocarcinoma noted in mice; C57BL IL10^−/− were less prone to develop colitis and carcinoma, Balbc IL10^−/− were intermediate in susceptibility and 129 IL10^−/− mice developed the most severe disease ⁷. Once identified, we found that while H. hepaticus infection could be demonstrated in a number of immunologically normal mice without colitis, it was strongly associated with chronic proliferative typhlitis, colitis, and rectal prolapse in several immunodeficient mouse strains housed at NCI and MIT ^{8, 9}. The specific pathogen-free health status of these colonies was verified to include lack of serum antibodies to murine viruses and negative cultures for respiratory pathogens, Salmonella sp., and Citrobacter rodentium. These findings suggested infection with H. hepaticus was strongly associated with the development of colitis in a number of laboratory mouse strains, and raised the question of whether H. hepaticus might be an etiologic agent of murine colitis.

Establishing the Link Between Helicobacter spp. and IBD

To establish the link between H. hepaticus and the development of colitis, we demonstrated that H. hepaticus inoculated into Helicobacter spp. free defined flora C.B. 17 scid mice reconstituted with CD45RB^high T cells resulted in severe IBD, similar to that noted with human disease ¹⁰. The expression of typhlitis and colitis was statistically less severe in defined flora scid mice that had been reconstituted with T cells but not infected with H. hepaticus when compared to reconstituted scid mice infected with H. hepaticus ¹⁰, demonstrating that infection with this single species could significantly exacerbate colitis in a murine disease model ^{11, 12}. Similar results were obtained following targeted infection of C57BL IL-10-deficient mice with H. hepaticus ^{13, 14}. Targeted infection experiments have now demonstrated that H. hepaticus is able to induce colitis and in some cases lower bowel carcinoma in a number of mouse strains with defects in immune function and/or regulation (Table 1). Interestingly, the same strain susceptibility noted by Berg et al in IL10^−/− mice was also documented in Rag^−/− mice developing colitis associated lower bowel carcinoma when experimentally infected with H. hepaticus ^{7, 15}.

Table 1.

H. hepaticus- and H. bilis-associated IBD in mice^a

Genetic status of mice	Type of defect	Pathology	Reference
CD45RB (high)-reconstituted ICR    defined flora scids	Reconstitution with naïve CD4+ T cell	Typhlocolitis	10
TCRα, B mutants	Abnormal T cell receptors	Typhlocolitis	88
Scid ICR-defined florab	Lack T and B cell	Typhlocolitis	89, 90
C57BL/IL-10−/−c	Knockout	Typhlocolitis	14, 18, 19, 36
129SvEv/Rag2−/−	Knockout	Typhlocolitis, colon cancer	28, 56
C57BLRag2−/−	Knockout	Typhlocolitis	14
IL-7−/−/RAG-2−/−	Double Knockout	None	27
A/JCr	Normal	Typhlitis	3
Swiss Webster gnotobiotic	Normal	Enterocolitis	91
129SvEv/NF-κβ (p50−/−p65+/-)	Double Knockout	Typhlocolitis	92
mdrla−/−d	Lack P-glycoprotein	Typhlocolitis	93
SMAD3−/−d	Knockout	Typhlocolitis, colon cancer	94

^aIn mice of the same genetic status which had H. hepaticus (or other Helicobacter spp.)-negative microflora, no intestinaldisease was noted.

^bMice infected with H. bilis also produced IBD⁸⁹.

^cIBD also produced in C57Bl/IL-10^−/− mice experimentally infected with a novel urease-negative Helicobacter sp.⁹⁵; now namedH. typhlonius ⁹⁶; also IBD produced with H. trogontum ⁹⁷ and H. cinaedi ⁷⁴.

^dH. bilis produces IBD⁹³, ⁹⁴, and colon cancer⁹⁴.

Basis of the GI inflammatory response to H. hepaticus

The experiments described indicate that H. hepaticus infection can induce colitis in certain susceptible mouse strains, but the nature of the inflammatory response and why certain strains were susceptible required further investigation. Seminal experiments from the early 1990’s demonstrated that mice lacking IL-10 developed spontaneous colitis, which was significantly less severe when mice were housed in SPF conditions, suggesting enhanced susceptibility to microbiota induced inflammation ¹⁶. In contrast to H. hepaticus monoassociated C57BL IL10^−/− mice which did not develop colitis, Kullberg and co-workers directly demonstrated that targeted infection of SPF housed C57BL IL-10 mice with H. hepaticus significantly exacerbated the development of colitis ^{13, 17}. The inflammatory response in the H. hepaticus-infected C57BL IL-10-deficient mice was characterized by a Th1-like cytokine profile ^{13, 18}. More recently these results have been replicated by demonstrating that wild-type (WT) C57BL mice treated with an antibody directed at the IL10 receptor (anti-IL-10R), but not those treated with a control antibody, develop colitis following infection with H. hepaticus ¹⁹. Colitis in anti-IL-10R treated mice is characterized by increased expression of mRNA coding for IFN-γ and IL-17 within the colon. Mesenteric lymph node cells (MLNs) derived from H. hepaticus-infected mice treated with anti-IL-10R secreted higher levels of IL-17 and IFN-γ compared to control mesenteric lymph node cells following exposure to soluble helicobacter antigen ¹⁹. Mice-deficient in IL-12p40 but not those deficient in IL-12p35 were resistant to H. hepaticus-induced colitis following anti-IL-10R treatment, which was associated with decreased expression of IFN-γ and IL-17, suggesting that IL-23 rather than IL-12p70 was necessary for the generation of both Th1 and Th17 responses observed following H. hepaticus infection. These experiments indicated that H. hepaticus can induce antigen-specific Th1 and Th17 responses in mice with impaired IL-10 signaling.

While H. hepaticus infection of mice with impaired IL-10 signaling induces CD4-mediated Th1 and Th17 responses, the exact role of both IFN-γ and IL-17 in H. hepaticus-mediated colitis has been more difficult to discern. Initial experiments with an IFN-γ depleting antibody suggested that IFN-γ was necessary for H. hepaticus to induce colitis in IL-10-deficient mice ¹³. However, the severity of H. hepaticus-induced colitis in mice doubly-deficient for IL-10 and IFN-γ was very similar to that in mice lacking IL-10 alone ¹⁸. Further, whereas H. hepaticus-induced inflammation within the colon of IFN-γ-deficient mice treated with anti-IL-10R was decreased compared to similarly treated WT mice, there was no difference in the level of inflammation in the cecum between genotypes ¹⁹. Although to our knowledge, the role of IL-17 in regulating H. hepaticus-induced inflammation in mice with impaired IL-10 signaling has not been directly tested, there are conflicting results from experiments testing the role of IL-17 in the development of colitis that develops follow the adoptive transfer of CD4⁺CD45RB^high cells into lymphocyte-deficient hosts ^{20, 21}. Further, as it has been suggested that other Th17 cell-derived cytokines such as Il-17F ²⁰ and IL-22 ^{22, 23} can modulate disease severity in T cell-dependent models of colitis, the role of these cytokines in H. hepaticus-dependent colitis requires additional study.

Although the exact role of T helper subsets in the development of H. hepaticus-induced colitis in susceptible hosts remains to be determined, increasing evidence has focused attention on the innate inflammatory response to H. hepaticus. Early studies demonstrated that H. hepaticus could induce colitis in certain strains of lymphocyte-deficient mice including scid and 129SvEv/Rag^{−/− 10, 24-27}, and progressed to colonic carcinoma with neoplastic peritoneal invasion in some murine models ^{24, 28}. H. hepaticus-induced innate colitis appeared to be dependent upon IL-23, as antibody depletion of both IL-12 p40 and IL-12 p19 (the components of IL-23) inhibited H. hepaticus-induced colitis in 129SvEv/Rag^-/- mice ²⁹. Antibody depletion assays have also demonstrated an important role for TNF in H. hepaticus-induced innate colitis ²⁵.

Surprisingly, even in the innate colitis model, H. hepaticus infection induced robust levels of the T helper associated cytokines IL-17 and IFN-γ within the colon of 129SvEv/Rag^−/− mice ²⁹, and treatment of H. hepaticus-infected 129SvEv/Rag^−/− mice with either IL-17 or IFN-γ depleting antibody ameliorated colitis ³⁰. This suggested that cells of the innate immune system expresses IL-17 and IFN within the colon following H. hepaticus infection. Indeed, in 129SvEv/Rag^−/− mice infected with H. hepaticus, there was a marked increase in the percentage of colonic lamina propria (LP) cells that stain positively for IL-17 alone or both IL-17 and IFN-γ following overnight stimulation with IL-23 ³⁰. These IL-17^± cells lacked lineage markers (CD11b, GR1, B220), but expressed high levels of Thy1 ³⁰, and thus bear similarities to a previous described rare population of lamina propria cells termed Lymphoid Tissue inducer-like (LTi-like) cells ^{31, 32}. Some LTi-like cells express CD4 on their cell surface ³¹, while these newly described IL-17⁺ innate lymphoid cells do not ³⁰. However they share a number of other cell surface markers including IL-7Rα ^{30, 31}. Interestingly, there was a marked increase in the frequency of Thy1^high innate lymphoid cells within the colonic lamina propia of 129SvEv/Rag^−/− following H. hepaticus infection, and when isolated by sorting, these cells produced IFN-γ, IL-17, and IL-22 after overnight culture, which was enhanced by treatment with IL-23 ³⁰.

It has previously been shown that the development of LTi and LTi-like cells, as well as classic Th17 cells require the transcription factor, RORγt ^31-33. Interestingly, H. hepaticus-induced innate⁻Thy1^high+ LP cells expressed higher levels of RORγt than remaining Thy1⁻ LP cells ³⁰, and both C57BL/6Rag^−/− mice treated with anti-Thy1 antibody, or C57BL/6Rag^−/− mice that lacked the gene for RORγt, were resistant to anti-CD40-induced innate colitis ³⁰. These data strongly suggest that Thy1⁺ RORγt-expressing innate lymphoid cell plays a critical role in driving intestinal inflammation. Further, if these innate lymphoid cells require the IL-7 receptor for their survival and/or expansion in similar fashion to LTi cells ³⁴, this could explain the earlier observation that H. hepaticus is unable to induce colitis in 129SvEv/Rag^−/− mice that also lack IL-7 ²⁷.

These studies argue that innate immune mechanisms play a critical role in the ability of H. hepaticus to induce colitis in susceptible hosts. However, it is important to recognize that H. hepaticus infection does not induce colitis in most WT mouse strains despite persistent infection, suggesting that a lymphocyte-mediated mechanism inhibits H. hepaticus-induced innate inflammation. Consistent with this, several groups have shown that the adoptive transfer of CD4⁺CD25⁺ T regulatory-cells into 129SvEv/Rag^−/− mice can inhibit the development of colitis following H. hepaticus infection ^{25, 28}. Inhibition depends upon the ability of transferred cells to express IL-10, as IL-10-deficient CD4⁺CD25⁺ T cells are unable to inhibit H. hepaticus-induced innate inflammation, and in fact appear to exacerbate disease ^{24, 25}. Remarkably, transfer of CD4⁺CD25⁺ T regulatory cells into H. hepaticus-infected 129SvEv/Rag^−/− mice with colitis leads to rapid resolution of disease ^{24, 35}. Several studies revealed that potency of CD4⁺CD25⁺ to protect against or treat IBD can be enhanced by H. hepaticus infection of donor mice ^{36, 37}. Thus CD4⁺CD25⁺ T regulatory cells can both inhibit and treat H. hepaticus-induced innate inflammation in the lower bowel, based on an IL-10-dependent mechanism. Interestingly, CD25⁺ cells are not the only CD4⁺ subtype with regulatory activity, as it has been demonstrated that CD4⁺CD45RB^lowCD25⁻ cells harvested from H. hepaticus infected C57BL/10 mice are potent inhibitors of colitis induced by adoptive transfer of IL-10-deficient CD4⁺ T cells into H. hepaticus-infected C57BL/10Rag^−/− mice ³⁶.

Experiments described above indicating that the recognition of H. hepaticus by the innate immune system plays a key role in regulating the ensuing inflammatory response in the lower bowel, suggest that understanding the innate inflammatory signaling pathways and receptors utilized by H. hepaticus could lead to important insights into the basis of microbiota-driven colonic inflammation. H. hepaticus challenge of ex vivo cultured bone marrow-derived macrophages induces activation of both the NF-κB and ERK pathways ^{26, 38}. Evidence for both pro and anti-inflammatory roles for individual NF-κB subunits in the response to H. hepaticus come from experiments demonstrating that 129SvEv/Rag^−/− mice that also lack the NF-κB subunit c-Rel are resistant to H. hepaticus-induced colonic inflammation ³⁹, while 129SvEv mice that lack the p50/p105 subunit of NF-κB are sensitized to H. hepaticus-induced colitis, based on a defect within the innate immune system ^{26, 40}.

Identifying the innate immune receptors responsible for inducing inflammatory signaling in response to Helicobacter spp. infection is an active area of reserach. Whole H. hepaticus bacteria are able to induce NF-κB activation in HEK293 cells transfected with TLR2 but not those transfected with TLR4 ⁴¹. However, a recent study demonstrated that the absence of TLR2 in 129SvEv/Rag^−/− mice does not influence the severity of the H. hepaticus-induced colitis ⁴². Nonetheless, there is strong evidence suggesting the involvement of TLRs in mediating the in vivo inflammatory response to H. hepaticus as irradiated 129SvEv/Rag^−/− reconstituted with Myd88 deficient bone marrow cells are resistant to H. hepaticus-induced innate colitis ⁴³.

Intracellular receptors of the Nod family plan an important role in recognizing bacterial products such as peptidoglycan that are injected by bacterial secretion systems⁴⁴. It has recently been reported that the genome of H. hepaticus codes for a type VI secretion system that may function to inject substrates into host epithelial cells ⁴⁵. Indeed, it has been suggested that the ability of H. muridarum, a murine EHS, to induce NF-kB reported activity in HEK293 cells depends upon Nod1 ⁴⁶. Although, of note, H. hepaticus is a very poor inducer of NF-kB activation in HEK293 cells ⁴⁶, and a weak inducer of inflammatory gene expression in the murine intestinal crypt epithelial cells (m-IC_cl2 cells) ⁴⁷, suggesting that H. hepaticus is unlikely to engage Nod1 in these epithelial cell lines. In contrast, it has been found that H. hepaticus, as well as LPS derived from H. hepaticus, can inhibit TLR4 induced inflammatory gene expression in m-IC_cl2 cells ⁴⁷, raising the possibility that H. hepaticus may have prominent anti-inflammatory properties on epithelial cells. An anti-inflammatory role for an H. hepaticus-derived product is supported by studies showing that an H. hepaticus mutant lacking the type VI secretion system colonized the intestine of C57BL/6RAG-/- recipients of CD4⁺CD45RB^hi T cells to higher levels than the WT H. hepaticus strain, and induced elevated levels of inflammatory gene expression ⁴⁵. Finally, a role for Nod2 in recognition of H. hepaticus has been suggested by studies demonstrating higher levels of H. hepaticus in infected Nod2-deficient C57BL/6 mice than in infected WT mice, although intestinal pathology was not detected following infection of either WT or Nod2-deficient mice ⁴⁸.

Helicobacter hepaticus infection modulates tumorigenesis at intestinal and extra-intestinal sites

At the time of initial characterization in A/JCr mice, it was recognized that in addition to active chronic hepatitis, infection with H. hepaticus was also associated with the development of hepatocellular carcinoma in A/JCr as well as live tumor development in AXB, B₆C₃ and a substrain of C57BL mice ^{1, 3, 49-52}. Previous studies using germ-free mice link chronic bacterial infection with chronic inflammatory diseases in the lower bowel ⁵³. Targeted bacterial infection with H. hepaticus in 129SvEv/Rag2^−/− mice lead to colitis-associated colorectal cancer (CRC) with malignancy invading the peritoneal cavity ^{24, 28}. Gut bacteria-induced lesions in 129SvEv/Rag2 -deficient mice resembled neoplasia in human patients ⁵⁴ with IBD, in that carcinoma clearly arises from colitis-associated dysplastic epithelial foci that become locally invasive ⁵⁵. The finding that 129SvEv/Rag2^−/− mice lacking lymphocytes exhibited CRC demonstrated that innate immunity is sufficient for development of carcinoma ²⁸. However, immunologically intact 129 wild type mice infected with H. hepaticus did not develop carcinoma, showing that IL-10-dependent activities of adaptive immune cells normally serve to reinforce homeostasis and prevent cancer in the lower bowel ^{28, 37}. Adoptive transfer studies using CD4⁺ cell subsets demonstrated that an imbalance between IL-6, IL-10 and Tgf-β increased risk of neoplastic epithelial invasion ³⁷. Interestingly, H. hepaticus-induced malignancy in the 129SvEv/Rag2^−/− mice was readily reversible by blocking underlying bacteria-driven inflammation with antibodies directed at TNF and IL-6 ^{37, 56}.

Pathogenic gut microbial infections have also been shown to trigger development of intestinal adenomatous polyps in mice with a mutated Apc gene (Apc^Min/+ or Min) ⁵⁷. Mice heterozygous for Apc gene (Apc^Min/+) develop a large number of intestinal polyps by 3 months of age ⁵⁸. This pre-neoplastic process has been shown to be inhibited by adoptive transfer of interleukin 10 (IL-10)–dependent T_Reg cells, and depends on their ability to produce IL-10 ^{59, 60}. Targeted infection with H. hepaticus increases multiplicity and malignancy of intestinal polyps in the Apc^Min/+ mouse model by an inflammation-associated TNF-α-dependent mechanism ⁵⁷. Likewise, infection with other pathogenic enteric microbiota including Citrobacter rodentium ⁶¹ and an enterotoxigenic Bacteroides fragilis ⁶² have been shown to promote colon tumorigenesis in Apc^Min/+ mice. Recent evidence points to key roles for IL-17 as well as IL-23 in colonic hyperplasia and tumor formation ^{62, 63}. Taken together these models may provide novel mechanistic insights in colon carcinogenesis in humans.

More recently, experiments examining the ability of GI tract bacteria and subpopulations of immune cells to induce or inhibit IBD have revealed surprising insights into extra-intestinal carcinogenesis. For example, H. hepaticus infection was found to trigger mammary tumorigenesis in Apc^Min/+ in the absence of IBD ^{57, 64}. Mammary tumors also developed rapidly in susceptible mice after transfer of CD4⁺ CD45RB^high T cells ^{57, 63}. While WT Treg completely inhibited H. hepaticus-induced diseases in these models, IL-10-deficient Treg cells exacerbated rather than inhibited H. hepaticus-induced extra-intestinal pathology ^{24, 37}. Interestingly, prior exposures of Treg donor mice to H. hepaticus enhanced their ability to inhibit extra-intestinal pathology ^{57, 63, 64}, raising the possibility that therapies aimed towards intestinal homeostasis may offer sustained protection from systemic diseases.

Virulence Factors of Enterohepatic Helicobacters

Human IBD as well as IBD in mouse models is multifactorial, with genetic and environmental factors contributing to pathogenesis. The complexity of the system necessitates parallel investigations into bacterial virulence factors, gene expression changes in intestinal tissue, host innate and adaptive immune responses, changes in the gut microbial flora, and of course, lesion development in the intestine. Using the H. hepaticus genome, we have begun studying in detail, with the use of isogenic mutants, how different virulence genes play a role in chronic intestinal inflammation. The pathogenic potential of several virulence factors from H. hepaticus have been well characterized.

Of the known EHS, H. hepaticus strain 3B1 (ATCC 51449), the protype of EHS, has been the most extensively investigated. This murine pathogen contains a circular genome of ~1.8 Mb with 1875 predicted open reading frames ⁶⁵. The G+C content of the H. hepaticus genome is 35.9% which places it between the G+C content of H. pylori (39%) and C. jejuni (30.5%). In addition, ~ 50% of predicted H. hepaticus genes are orthologous to those in the genome of either H. pylori or C. jejuni, both of which cause significant gastrointestinal diseases in humans ⁶⁵.

Cytolethal distending toxin (CDT)

Bacterial CDTs are present in multiple gram negative pathogens including strains of Salmonella spp, Shigella spp, C. jejuni, and E. coli, and belong to the AB₂-type of toxins. In general these toxins are comprised of subunits CdtA and CdtC which together bind subunit CdtB and deliver CdtB, the active subunit of the toxin, intracellularly ⁶⁶. Several biological functions of the CdtB subunit have been reported, including a DNase I-like function, cell cycle arrest, phosphatase activity and eventual cell death ⁶⁷. The cdtB orthologue and CDT activity are found in several enterohepatic helicobacters, including H. hepaticus, H. bilis, H. canis, and H. marmotae as well as in H. cineadi and H. pullorum isolated from humans ^68-70.

H. hepaticus 3B1 and an isogenic mutant of the cytolethal distending toxin gene (cdtB-deficient HhcdtBm7) were evaluated in outbred Swiss Webster mice. Whereas inoculation with WT 3B1 lead to persistent colonization, colonization by the mutant failed in female mice and was transient in males ⁷¹. Wildtype 3B1 infection caused significantly increased expression of ileal and cecal IFN-γ mRNA, higher Th1-induced IgG2a responses, and increased mucosal IgA secretion compared to the isogenic mutant ^{13, 71}. CDT-deficient isogenic H. hepaticus mutants also lose their ability to persistently colonize the large intestine of B6.129-IL10^tm1Cgn (IL10^−/−) mice. Importantly, the mutant-infected mice developed significantly less severe typhlocolitis when compared to WT 3B1 infected mice ^71-73. Mutation of the CDT gene of H. cinaedi, a human pathogen, also reduced its ability to induce lower bowel pathology in IL-10-deficient mice ⁷⁴.

Molecular mechanisms operable in H. hepaticus CDT induced cell death are also being actively investigated. CDT activated the ATM-dependent DNA damage check point response in a myc-dependent fashion ⁷⁵. In addition, exposure of INT407 cells to H. hepaticus CDT resulted in upregulation or pro-apoptotic Bax, release of Cyt c, and activation of caspases 3/7 and 9 ⁷⁶. Events operable in H. hepaticus CDT-induced cell arrest and apoptosis are summarized in Figure 1.

Figure 1.

ATM-dependent checkpoint responses and apoptosis triggered by H. hepaticus CDT-induced DNA damage ^{75, 76}. DSB, double strand breakage; Nbs 1, Nijmegen Breakage Syndrome 1 protein; ATM, Ataxia Telangiectasia Mutated; CHK2, checkpoint kinase 2; Bcl2, B-cell leukemia/lymphoma 2; Bax, Bcl-2–associated X protein; Cyt c, Cytochrome c; Apaf-1, apoptotic peptidase activating factor 1.

H. hepaticus pathogenicity island (PAI)

PAIs are distinct genetic elements on bacterial chromosomes that encode one or more virulence-associated factors and often have a G+C content distinct from the rest of the core genome ⁶⁵. Several PAIs, such as LEE in enteropathogenic Escherichia coli, cag in Helicobacter pylori and SaPl1 in Staphylococcus aureus, significantly contribute to pathogenicity of these infectious agents ⁷⁷. In the genome of H. hepaticus 3B1, a 71-kb region, HHGI1, displays genetic features characteristic of PAIs ⁶⁵. HHGI1 is highly variable among certain H. hepaticus isolates. Male A/JCr mice infected with H. hepaticus strain MIT 96-1809 lacking the entire HHGI1 or strain MIT 96-284 missing ~62 out of 71-kb of HHGI1 developed less-severe hepatitis than those infected with H. hepaticus 3B1 which has an intact HHGI1 ⁷⁸. An isogenic mutant (HhPAId1) of H. hepaticus 3B1 containing a deletion of 19 predicted genes (HH0250 to HH0268) within HHGI1 colonized the intestine of B6.129-IL10^tm1Cgn (IL10^−/−) mice at higher levels but provoked less severe typhlocolitis and hyperplasia when compared to H. hepaticus 3B1 ⁷⁹. The magnitude of the Th1-associated IgG2c response against this isogenic mutant (HhPAId1) was less than against wildtype H. hepaticus and was accompanied by suppressed cecal and colonic mRNA expression levels for proinflammatory IFN-γ, TNF-α and IL17a. These results demonstrate that HHGI1 contributes to the pathogenicity of H. hepaticus at least in part via up-regulation of proinflammatory mediators.

Recently, a study indicated that HHGI1 encodes for a set of 12 genes (HH0237, HH0242 to HH0245, HH0247 to HH0252, HH0291) which are homologous to known bacterial type VI secretion systems (T6SS) ⁴⁵. T cell-reconstituted Rag1^−/− mice infected with a T6SS-deficient ΔIcmF H. hepaticus mutant (encoded by HH0252) developed higher colonization levels of the mutant compared to wildtype H. hepaticus-infected mice, consistent with higher levels of mutant HhPAId1 cecal colonization ⁷⁹. In addition, infection with ΔIcmF H. hepaticus induced higher expression in the colon of proinflammatory mediators including IL-17, IL-23, IL-1β, TNFα and IFNγ than infection with WT H. hepaticus 3B1 ⁴⁵. However, unlike the HhPAId1 mutant, the T6SS mutant ΔIcmF, and H. hepaticus 3B1 induced a similar degree of colitis by 2-4 weeks postinoculation ^{45, 79}. Therefore, further investigations will be required to define the role(s) of T6SS and additional components in HHGI1 in the pathogenesis of H. hepaticus.

Urease

Gastric helicobacters produce urease which is essential for establishing persistent colonization in the acidic milieu of the stomach ⁸⁰. Approximately 40% of known EHS species, including H. hepaticus, are urease-positive ⁸¹. The H. hepaticus ure gene cluster is similar to that of H. pylori; however, regulatory mechanisms of urease activity in H. hepaticus are distinct from those in H. pylori. For example, H. hepaticus urease does not confer acid-resistance, transcription of ure genes is not induced by acid, and urease activity is suppressed by the transcriptional regulator Fur ^{82, 83}.

By infecting male A/JCr mice with a urease-deficient, isogenic mutant (HhureNT9) of H. hepaticus 3B1, we demonstrated that the lack of urease activity did not affect the level of cecal colonization but its presence was essential for hepatic colonization ⁸¹. Consequently, the HhureNT9-infected mice developed less severe hepatitis and produced significantly lower hepatic mRNA levels of proinflammatory cytokines IFN-γ and TNF-α when compared to the H. hepaticus 3B1-infected mice ⁸¹.

Enzymes in oxidative stress resistance

To survive and establish colonization in the host, bacteria have evolved various mechanisms to combat oxidative stress that confers bactericidal effects via reactive oxygen species. The H. hepaticus catalase enzyme KatA, which converts hydrogen peroxide into molecular oxygen and water, plays a key role in this process. The amino acid sequence of H. hepaticus KatA contains three motifs (R-F-Y-D, RERIPER and VVHAKG) in the haem-ligand domain and three surface-predicted motifs that are shared among bacterial and mammalian catalases ⁸⁴. KatA-deficient H. hepaticus are more sensitive to exposure of oxygen and H₂O₂ and also suffer severe DNA fragmentation by H₂O₂ treatment compared to H. hepaticus 3B1 ⁸⁵.

Bacterial hydrogenase, a hydrogen uptake hydrogenase, oxidizes molecular hydrogen to yield protons and electrons which have a key role in energy conservation in bacteria. H. hepaticus genes or HH0056 to HH0059 encode subunits of hydrogenase HyaA to HyaD. Mutation in the H. hepaticus hyaB (coding for the large subnit) does not affect cecal and hepatic colonization levels when compared to H. hepaticus 3B1 in male A/JCr mice; however, liver pathology is diminished in the hyaB mutant-infected mice ⁸⁶. It is unknown how hydrogenase contributes to the pathogenesis of H. hepaticus, but it is speculated that energy released by hydrogen oxidation may be utilized to promote synthesis of virulence-related proteins or enzymes that result in hepatic or intestinal inflammation and necrosis ⁸⁶.

Flagellar genes

All Helicobacter species possess flagella for motility; H. hepaticus is a spiral rod with bipolar sheathed flagella. In other related bacteria, flagellar genes are mainly regulated by the sigma factor FliA (σ28). H. hepaticus contains two identical copies of the gene encoding the major flagellin subunit FlaA (open reading frames HH1364 [flaA1] and HH1653 [flaA2]) ⁶⁵, which is regulated by the sigma factor FliA (σ28). Mutations in fliA or both copies of the flaA gene resulted in no detectable synthesis of FlaA as well as severely truncated flagella, thus both mutants were nonmotile and unable to colonize mice ⁸⁷. The flaA_1 H. hepaticus mutant, which has flagella but decreased motility, also does not colonize mice, indicating that full motility is a prerequisite for intestinal colonization by H. hepaticus, and that the presence of flagella alone is not sufficient.

Summary

As described in this review, Helicobacter hepaticus has played a central role in shaping our understanding of how intestinal microbiota interact with the host to produce both inflammatory and tolerant responses. The study of immunopathology that develops following H. hepaticus infection of susceptible mice has helped to define the importance of key inflammatory mediators including IL-23 and IL-17, and highlighted the importance of the innate immune response in the development of chronic intestinal inflammation. The use of H. hepaticus infection models has provided definitive evidence that regulatory T cells directly inhibit the innate inflammatory response to lower bowel microbiota and delineated the critical role of IL-10 in this process. Remarkably, H. hepaticus infection has powerful effects on carcinogenesis both within the intestine and at extraintestinal sites, and many of the same immunoregulatory pathways that influence the gastrointestinal inflammatory response to H. hepaticus also appear to influence the development of cancer. The complete genomic sequence of H. hepaticus has advanced the identification of potential virulence factors, and evaluating the function of these factors is ongoing. In addition, genome sequencing of five additional EHS isolated from humans with gastrointestinal disease (H. bilis, H. canadensis, H. cinaedi, H. pullorum and H. winghamensis) has been completed and these results can be found at http://www.broadinstitute.org/annotation/genome/Helicobacter_group/. Comparative genome analysis of these 5 genomes plus H. hepaticus will provide insight into the etiopathogenesis of EHS infection in humans. For example, preliminary data indicates that both H. cinaedi and H. pullorum have the existence of PAI’s with similar features to the HHGI found in H. hepaticus. Studies are needed to ascertain whether these PAI’s impart similar in vivo characteristics in humans. Additional areas for future investigation include defining the cellular and molecular pathways that contribute to host recognition of H. hepaticus, and determining how initial recognition events lead to activation of innate and acquired inflammatory effector mechanisms. Further, clarifying how responses to H. hepaticus infection influences carcinogenesis at local and distant sites could lead to important new insights into the relationship between specific microbiota and the development of cancer. Finally, delineating the targets of H. hepaticus associated virulence factors could lead to novel approaches to treating microbial induced inflammation in the lower bowel.